Various conditions in optical communication systems make it desirable to be able to selectively attenuate one or more particular wavelengths in an optical signal relative to other wavelengths in that signal. For example, it may be advantageous to provide a flat gain response across multiple wavelength channels of an optical signal. This typically requires providing separate attenuation circuitry for each wavelength to be attenuated. Although conventional systems exist for providing attenuation in optical signals, no system has emerged that provides cost effective attenuation in multiple wavelength systems. This problem becomes increasingly acute as optical systems strive to implement more and more wavelength channels.
One particular problem that can arise in an optical communication system supporting many wavelengths involves controlling the gain tilt in the transmitted signal. Existing optical communication systems have typically been limited to using the conventional (“C”) band of wavelengths to communicate optical signals. With the increasing demand for bandwidth, the capacity of communication systems is being expanded by the addition of new communication bands. For example, future communications systems will likely use the long wavelength (“L”) band and possibly even the short wavelength (“S”) band.
As additional wavelength bands are utilized and the net power in the fiber is increased, a problem can arise from an inter-channel Raman effect. In particular, longer wavelength channels can rob power from the shorter wavelength channels, creating a gain tilt after propagation through the fiber. The gain tilt can become increasingly pronounced as links of amplified fiber segments are cascaded.